Literature Lab: a method of automated literature interrogation to infer biology from microarray analysis
© Febbo et al; licensee BioMed Central Ltd. 2007
Received: 12 July 2007
Accepted: 18 December 2007
Published: 18 December 2007
The biomedical literature is a rich source of associative information but too vast for complete manual review. We have developed an automated method of literature interrogation called "Literature Lab" that identifies and ranks associations existing in the literature between gene sets, such as those derived from microarray experiments, and curated sets of key terms (i.e. pathway names, medical subject heading (MeSH) terms, etc).
Literature Lab was developed using differentially expressed gene sets from three previously published cancer experiments and tested on a fourth, novel gene set. When applied to the genesets from the published data including an in vitro experiment, an in vivo mouse experiment, and an experiment with human tumor samples, Literature Lab correctly identified known biological processes occurring within each experiment. When applied to a novel set of genes differentially expressed between locally invasive and metastatic prostate cancer, Literature Lab identified a strong association between the pathway term "FOSB" and genes with increased expression in metastatic prostate cancer. Immunohistochemistry subsequently confirmed increased nuclear FOSB staining in metastatic compared to locally invasive prostate cancers.
This work demonstrates that Literature Lab can discover key biological processes by identifying meritorious associations between experimentally derived gene sets and key terms within the biomedical literature.
The accelerating expansion of biomedical research outpaces most individual attempts at comprehensive review even in relatively narrow fields. Just as the vast sequence data available for the human [1, 2] and additional organisms [3–5] require sophisticated genomic browsing tools [6–8], computational methods are required to thoroughly explore the corpus of biomedical literature. Many computational methods for interrogating the scientific literature have been developed . These programs can be broadly defined as methods for information retrieval and those for information extraction .
Existing methods can identify significant association between individual genes and terms from the medical subject heading (MeSH) index and Gene ontology (GO) databases , manually curated biological lists , or disease-specific lists . These prior methods demonstrate that genes with disease-specific differential expression can be strongly correlated with key terms within the medical literature [13, 14]. In addition, gene-gene associations within the literature have been combined with multiple available databases to extend associations beyond the literature alone .
While these and similar approaches have underscored the potential of automated literature searching to facilitate discovery, few have provided both methods for assessing the statistical strength of identified associations and supported their methods with experimental validation. Here, we describe and apply "Literature Lab", a method of automated data retrieval confined to publicly available citations and abstracts. Literature Lab statistically assesses identified associations within the corpus of medical literature between sets of experimentally derived genes and key terms derived from curated or MeSH lists. We demonstrate that our methodology can identify previously reported relationships and can result in discovery.
Gene Expression Sets
Literature lab was applied to three gene sets derived from previously published microarray data and one gene set from as yet, unpublished dataset to determine if literature mining identified important metabolic, physiologic, or pathway activity. These gene sets include, 1) the top 100 genes with increased expression during in vitro exposure of human leukemia cells (HL60, AML cell line) to ATRA ("UCT"), 2) the genes differentially regulated with a transgenic model of prostate neoplasia (MPAKT) following exposure to RAD001, 3) the 70 genes used to predict outcome in localized node-negative breast cancer, and 4) genes differentially expressed between malignant epithelial cells in local v. metastatic prostate cancers. The genes lists are included in additional files and a full description of the development of each gene set is in the supplemental methods (see Additional File 1).
Protein from fresh ventral prostates were extracted in RIPA buffer [10 mM sodium phosphate (pH7.2) 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 1 mM NaOVa, 1 mM DTT, 5 mM NaF, 0.1% sodium deoxycholate, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 1 mM PMSF] separated by gel electrophoresis and transferred to nitrocellulose membrane (0.45 μM) as described[16, 17]. Membranes were blotted with anti-Hif-1α (kindly provided by J. Pouyssegur) and anti-tubulin (B-5-1-2) (Sigma) (1:1000). Blots were scanned and intensities were measured.
A prostate tissue microarray containing samples of benign prostate epithelium (n = 14), locally invasive prostate adenocarcinoma (n = 20), and metastatic prostate adenocarcinoma (n = 22) were stained for FOSB. The FOSB staining was performed as previously described [18, 19]. Briefly, 5 micrometer sections were cut from the TMA block, deparaffinized, rehydrated and subjected to microwaving in 10-mM Citrate buffer (pH 6.00) in a 750W oven, for 15 minutes. The polyclonal anti-FOSB primary antibody (Santa Cruz Biotechnologies, Inc.), was incubated (1:50 dilution) at room temperature in an automated stainer (Optimax Plus 2. 0; Biogenex, San Ramon, CA). Antigen-antibody reaction was revealed with standardized development times, using the Streptavidin method with 3, 3 diaminobenzidine as substrate. Meyer Hematoxylin was used as nuclear counterstaining. FOSB nuclear positivity was scored on a total of 50 nuclei per sample. The statistical difference in FOSB staining between the unpaired populations was assessed using a two-tailed Mann Whitney test (GraphPad Prism4 Software).
Application of Established Literature Mining Tools
MeSHer data for the gene lists was obtained from the website  by supplying the Affymetrix U133Plus2 Ids corresponding to the list of genes using a p-value threshold of 1.0 and no p-value correction.
GOMiner data for the gene lists was obtained by following the instructions for downloading (build 148) and installing the GOMiner application and SQL database from . A list of gene symbols for genes in the Affymetrix U133Plus2 probe set list was used for the 'Total' file and lists of symbols for the experimental genes used for the 'Changed' file. Other GOMiner settings used were the default values defined in the GOMiner application.
To highlight and prioritize cellular physiology, metabolism, or pathways differentially active within a microarray experiment, Literature Lab uses an experimentally derived gene list, pre-defined sets of non-ambiguous key terms ("domains"), inclusive gene nomenclature, comprehensive literature interrogation, and a comparison of the experimental gene set results with randomly generated gene sets. For each experimentally derived gene set, Literature Lab performs an automated literature search to determine the number of abstracts listing any of the identified genes and each term within the term list (or domain). Specific domains include those for cell metabolism (MeSH headings), cell physiology (MeSH headings), and cellular pathways (manually curated from multiple sources) (Additional File 2). Terms are ranked with a measure of association between genes and terms known as the product of frequency (PF) (see Additional File 1). The strength of the relationship between an experimental gene set and a term is determined by comparing the log (base 10) of the sum of the PF values for the experimental gene set (called the LPF) with a distribution of the same values for 1000 random gene sets of the same size as the experimental gene set. A score representing the number of standard deviations above or below the mean value for the random gene sets is assigned to each term and terms are ranked in decreasing order of this score. To highlight terms with particularly strong association with the gene list, we developed heuristic rules for labeling the relationships as "Strong" or "Moderate" (see Additional File 1 for Details) (the software used for these analyses is free for non-commercial use and available, see Availability and Results section).
Overall Architecture of the Software Implementation
• A series of programs which query PubMed for each term and each gene in the Gene Thesaurus using NCBI's Entrez Programming Utilities  and Sun's JAXB XML Binding classes . The result for each term or gene is a file containing a list of the PubMed Ids for the term or gene. In addition, summary files containing counts of the number of abstracts for each gene and for each term are prepared.
• A series of programs which identify the PubMed Ids in common for each term/gene combination from the lists of ids for each term and each gene. The results of these programs are files containing the PubMed Ids for each term/gene combination and files containing summary counts of the number of abstracts for each term/gene combination.
• A series of programs for preparing the analysis of a specific gene list. These programs compute the described statistics for the gene list and for one thousand random gene sets of the same size as the specific gene list being processed. The final step of this process prepares a Microsoft® Excel spreadsheet containing the results of the analysis. The Apache Software Foundation's POI classes (Java API to Access Microsoft® Format Files  are used to format the data in a form which can be viewed in Microsoft® Excel.
Gene annotation began by obtaining gene names, symbols, and "aliases" from Stanford University's SOURCE web site . Subsequent Boolean PubMed searches and manual review for precision (i.e. the ability to accurately identify abstracts related to the gene of interest) were used to develop a set of rules required for more specific automated searching (Additional File 1). In subsequent literature searches, gene terms consisting entirely of numerals, of three characters or less, or that were identified as excessively ambiguous through manual review were excluded. Algorithm-based disambiguation was not used . While the database of gene terms (referred to as the gene thesaurus) is frequently updated, for all experiments herein described, the gene thesaurus updated last on 12/30/2003 was used for all experiments presented (Additional File 3). 514 of the gene terms were specifically excluded based upon the manual review (Additional File 4). All gene term curation was performed prior to testing for associations.
Topic List Curation
Topic lists are sets of terms, each of which is to be tested for significant associations with the experimentally derived gene set. Topic lists used in the experiments were defined using terms from the MeSH Thesaurus provided by the National Library of Medicine. The list of terms for each topic set consisted of all terms and all descendents (with few exceptions) as listed in MeSH. The topic sets (presented as "name [Mesh ID]") used in these experiments were Cell Metabolism [G06.535] and Cell Physiology [G04.335] (Additional File 2). A few of the descendents of these MeSH terms were ignored because very few abstracts were associated with the term. When using MeSH terms to search PubMed, the search syntax used "MeSH Subheading Explosion" so that the resulting list of abstracts included abstracts coded with descendents (if any) of the MeSH term as well as abstracts coded with the MeSH term itself.
In addition, a topic set entitled "Pathways" was derived using the pathway descriptions from BioCarta  with some additional curation and testing (see Additional File 1 for methods and Additional File 2 for list of pathways). All pathway term curation was performed prior to testing for associations.
A list of the PubMed abstract ID's for each gene and each topic were obtained by searching PubMed using the appropriate terms and a constant date range. The PubMed Electronic Date (EDAT) was used to query a constant subset of the PubMed abstracts between 12/31/93 and 12/31/03. Whole text information was not used due to its incomplete and inconsistent availability. To determine if results changed over time with a fixed chronological time frame and pre-set search parameters, three gene lists ("Ideal", "UCT", and "Van't Veer") and three curated term lists (metabolism, physiology, pathway) were run repetitively every month for 5 months using the fixed dates above in order to assess if ongoing efforts at the NCBI to improve the search engine or annotation of the literature significantly influence our approach to literature mining.
Each topic list was analyzed independently. Within each topic set, specific terms were ranked according to how many times the term was associated with any of the genes in the gene list with respect to the total number of times the term is present in PubMed. We applied two different methods to measure the degree of intersection between any term and the gene set. First, we calculated the ratio of the number of abstracts containing any gene from the gene list AND the term divided by all abstracts containing the term (referred to as "expected abstracts"). The second score (called "product of frequency") takes into account the number of abstract mentions against the entire target gene set (see Additional File 1). In order to rank terms, we compared each terms score (by either method) to a distribution of scores between the same term and randomly selected gene sets. In the case of the product of frequency, the log(PF) (LPF) more closely approximated a normal distribution thus justifying the mean and standard deviation statistics used (see Additional File 1).
Testing against random gene sets
In order to provide a metric by which to interpret the rankings and determine the likelihood of finding a match given no association, we measure each topic score given the experimental set of genes against the distribution of scores from sets of genes chosen at random. Scoring 1,000 such random sets against the topic set, we obtain estimates for the mean and standard deviation of the F(geneset, topic) score for each topic. We tested if there was a significant difference in the statistics generated using or 1000 random sets of genes containing the same number of genes as the experimental gene set.
We applied Literature Lab to lists of genes generated through microarray experiments to determine if such an approach provides biological insight. We started with 3 lists of genes from previously published microarray experiments, one generated in vitro (HL60 cells treated with ATRA , one generated in vivo (MPAKT transgenic mice treated with RAD001 ), and one generated from human tumors samples (A 70 gene model of breast recurrence [31, 32]) to see if Literature Lab would correctly identify known biological processes and to how altering specific variables within Literature Lab impacts the results. Finally, we tested our method to a set of genes differentially expressed between local and metastatic prostate tumors and used immunohistochemistry to confirm the lead candidate.
Literature Lab associates "respiratory burst" with ATRA treatment of a leukaemia cell line
We tested the impact of gene set size and literature time frame on the successful association between "respiratory burst" and the genes with increased expression following ATRA. While the specific literature time frame had relatively little impact on the association (i.e. first or last 5 years of the 10 year period) (see Additional File 5), the number of differentially expressed genes included in the analysis did impact our results (see Additional File 6) Specifically, the association between "respiratory burst" and the experimentally derived list of differentially expressed genes did not stabilize until greater than 150 genes were included. While there are no obvious rules to guide the upper limit with respect to the number of genes to be used in literature lab, the variability of results observed with shorter gene lists suggests that gene lists numbering less than 25 will not provide robust results and for experiments comparing phenotypes with marked differences, gene lists greater than 150 are encouraged.
"Hif and Hypoxia" strongly associated with mTOR inhibition
"Matrix metalloproteinase 9" and "VEGF" are associated with Breast Cancer Prognosis
Primary tumors perhaps provide the biggest challenge for analysis because of the additional associated technical and biological variation. We applied Literature Lab to the set of 70 genes associated with outcome in two seminal papers predicting breast cancer recurrence using microarray data [31, 32]. Interestingly, of the cellular pathways investigated, "matrix metalloproteinase" and "VEGF" were strongly associated with the gene set (see Additional File 7). This unbiased association is strongly supported by prior studies finding MMP [35, 36] and VEGF  activity strongly associated with the recurrent phenotype  and supports Literature Lab's applicability to data from human samples.
Median of Absolute Value of Percentage Change in LPF over 5 Months
MPAKT Gene Set
HL60 Gene Set
vant Veer Gene Set
By identifying the abstracts lost or gained between sequential runs, we identified a number of causal factors for literature drift:
• Changes to the various components of the NCBI search engine that result in different results for the same query. In particular, changes to the PubMed "phrase dictionary" are frequent and can yield different results for the same query at different points in time.
• The assignment of MeSH terms to abstracts subsequent to the addition of abstracts to the PubMed database.
• The editing of PubMed abstracts so as to change the title or text.
• The occasional deletion of abstracts from the database. Many of these deletions appear to be the removal of duplicates added to the PubMed database in error.
Discovery of increased FOSB in Metastatic Prostate Cancer
Comparison with GOMiner, MESHER and GeneCite
To determine how Literature Lab compares with existing, publicly available sample annotation and literature mining technologies, we imported the AKT mouse and metastatic prostate cancer gene lists into GOMiner , MeSHer , and GeneCite . When the AKT mouse gene set was imported into GOMiner, 222 GO terms were found to be significantly associated (p ≤ 0.05) with the genes differentially expressed (see Additional File 8). Among the significant GO terms were "response to hypoxia" and "glycolysis" thus demonstrating that GOMiner is able to identify similar underlying biology but these terms were among a large number of significant terms and not prominent. Here we note that GOMiner and Literature Lab represent different approaches (enrichment analysis v. Literature mining, respectively) and, as such, the results are not directly comparable. However, the purposes of Literature Lab and GOMiner are similar (i.e. the identification of biological processes implicated by differential gene expression) and our results suggest that direct literature mining and our statistical approach provide insight that cannot be fully reproduced using GOMiner.
MESHER did not find a significant association between "Hypoxia", "HIF", and "Hypoxia p53" and the AKT mouse gene list and there was very poor comparison between MESHER and Literature Lab despite each method using MeSH terms (see Additional File 9). GeneCite identifies the abstracts associating the terms and genes but has no direct measure of significance other than the number of abstracts (which favours more general terms over more specific terms). In addition, GeneCite has lower precision and recall when compared to Literature Lab due to the lack of a thesaurus for gene nomenclature (see Additional File 10). For example, many of the abstracts for the CAT gene refer to felines and CAT scans and not to the CAT gene.
When the genes differentially expressed between local and metastatic prostate cancers were imported into GOMiner and MESHER, there was poor overlap with both. FOSB, the term significantly associated with metastatic prostate cancer and subsequently validated by immunohistochemistry, was either not present in the library (GOMiner) or not associated with the gene list (MESHER, "Biogenetics-MeSH – "Genes, fos"). Results for GeneCite (see Additional File 11) exhibit the same limitations previously described for the AKT mouse gene set.
Full utilization of publicly available, data-rich resources remains a universal challenge in contemporary scientific investigation. As technologies have diminished the cost and time associated with data collection, content within diverse repositories of data have increased exponentially. The medical literature is one such data repository and a repository that continues to grow rapidly. While investigators frequently use computational tools to interrogate genomic or gene expression data repositories, few use similar tools when reviewing the literature.
Literature Lab represents a method to comprehensively interrogate the literature for associations between a list of genes and a list of key terms in an unbiased manner in order to highlight potentially important biological processes implicated by the gene list. While there are many methods by which to develop a gene list, we have designed Literature Lab to aid in the analysis of microarray experiments which frequently associate the expression of hundreds to thousands of seemingly unrelated genes with cellular behaviors, in vivo phenotypes, or disease outcomes. We developed and refined our methodology using gene sets from previously published work and successfully tested Literature Lab on a novel dataset. The pathway term FOSB was ranked highest by Literature Lab and highlighted as having a "strong" association; an increase in nuclear FOSB staining was subsequently confirmed with immunohistochemistry.
Literature Lab is complementary to the increasingly prevalent pathway oriented approaches to the analysis of microarray data (Reviewed in ). As a general approach, these methods look for significant differential expression within a microarray experiment using pre-determined aggregations of genes (alternatively called gene sets, metagenes, or gene modules) rather than individual genes . Successful gene sets can identify underlying genetic abnormalities or signal transduction networks driving disease pathologies and help effectively bridge microarray data with biological significance [41, 42].
Some pathway approaches methods use the literature and publicly available annotations (Gene Ontology) to develop gene sets and use these gene sets to interrogate expression data [43, 44]. Literature Lab offers the opportunity to use a gene set derived from microarray data to interrogate the biomedical literature without a priori classification or annotation. As such, Literature Lab can appropriately interrogate the literature as it grows and evolves. When compared to two publicly available methods of analysis (GOMiner and MeSHer), the results of Literature Lab were more comparable with GOMiner. However, the statistical evaluation of associations identified by Literature Lab help improve the specificity of findings (highlighting strong associations) while maintaining sensitivity (neither GOMiner nor MeSHer identified the association between FOSB and genes with differential expression between local and metastatic prostate cancer). It should be noted, however, that given the difference in the approaches, our results cannot be interpreted as demonstrating the superiority of Literature Lab over GOMiner
Literature Lab remains dependent on the strength of the term lists and while we have demonstrated the use of lists for metabolism, physiology, and pathways, further development is focused on creating lists to include disease, pharmacological agents, drug toxicities, and many additional classes.
We initially anticipated that fixing the chronological interval for a query would ensure exact reproduction of the results. However, we identified literature drift within fixed retrospective intervals. While the degree of literature drift seems to range from minimal to moderate depending on the specific gene list, Literature Lab successfully limits the effects of literature drift especially for associations identified as "strong" with the current heuristics. Thus, while literature drift is unlikely to have significant impact on the associations identified by Literature Lab, some variation in the specific weights and rankings of associations will change even when investigators define a fixed chronological interval within which they perform their query.
For this initial description, we focused on developing a robust measure of association, a relatively useful measure of significance, and heuristic rules to highlight the most important associations. Clearly, the specifics of our methods will be the subject of further investigation and refinement. While we have identified some critical elements of success (avoiding measures of association that are driven by single gene-term associations and having a gene set size of 25 or more), work is ongoing to explore the effects of refining the genes based upon the statistical association between their expression and the phenotype, limiting Literature Lab to specific journals of high quality content, and increasing the number of sets of key terms with which to test the association between gene expression.
The methodology herein described for Literature Lab highlights the biomedical literature as a content-rich resource amenable to automated, comprehensive interrogation. As with most complex data, successful comprehensive interrogation requires filtering out the noise and finding valuable information. Our current methods of gene annotation, key term curation, and literature interrogation, can find strong associations and are likely to benefit a diverse scientific community.
Availability and Requirement
The instructions, software, and data required to perform an analysis of a gene list using the techniques described herein can be obtained from http://www.acumenta.com/freeware/instructions.html. Sun's Java Runtime environment Version 1.4 or higher is required in order to run the software (and may be downloaded from ). The software runs on both Windows platforms (Windows 2000 and later) and Linux platforms. Memory of 1 GB and 5 GB of available disk space are required (much of the disk space requirement is for temporary storage during the analysis). In addition, the results of the analysis are presented in a Microsoft Excel spreadsheet for viewing on any system having a spreadsheet viewer capable of rendering the Microsoft Excel format.
• Project name: Literature Lab
• Project home page: http://www.acumenta.com/freeware/instructions.html
• Operating system(s): Windows (2000 +) or Linux
• Programming language: e.g. Java
• Other requirements: Java Runtime environment 1.4 or higher
• License: Not required for Academic users
• Any restrictions to use by non-academics: License needed from Acumenta
P.G.F is a Damon Runyon-Lily Clinical Investigator (Grant #CA-29-05) and, in addition, was supported by NCI grants CA89031 and CA123175 during the design, analysis, and preparation of this manuscript. MGM, DAS, PRM, and SCT were supported by Acumenta, Inc during the design, analysis, and preparation of this manuscript. All analysis of data by investigators supported by Acumenta was performed in a blinded fashion; Acumenta supported investigators were unaware of the experiments from which the gene lists uploaded into Literature Lab were derived. KS was supported by the Department of Pediatric Oncology during the design and analysis of this work. DDV and ML were supported by the Department of Pathology at Dana Farber Cancer Institute during the design and analysis of this work.
- Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, Raymond C, Rosetti M, Santos R, Sheridan A, Sougnez C, Stange-Thomann N, Stojanovic N, Subramanian A, Wyman D, Rogers J, Sulston J, Ainscough R, Beck S, Bentley D, Burton J, Clee C, Carter N, Coulson A, Deadman R, Deloukas P, Dunham A, Dunham I, Durbin R, French L, Grafham D, Gregory S, Hubbard T, Humphray S, Hunt A, Jones M, Lloyd C, McMurray A, Matthews L, Mercer S, Milne S, Mullikin JC, Mungall A, Plumb R, Ross M, Shownkeen R, Sims S, Waterston RH, Wilson RK, Hillier LW, McPherson JD, Marra MA, Mardis ER, Fulton LA, Chinwalla AT, Pepin KH, Gish WR, Chissoe SL, Wendl MC, Delehaunty KD, Miner TL, Delehaunty A, Kramer JB, Cook LL, Fulton RS, Johnson DL, Minx PJ, Clifton SW, Hawkins T, Branscomb E, Predki P, Richardson P, Wenning S, Slezak T, Doggett N, Cheng JF, Olsen A, Lucas S, Elkin C, Uberbacher E, Frazier M, Gibbs RA, Muzny DM, Scherer SE, Bouck JB, Sodergren EJ, Worley KC, Rives CM, Gorrell JH, Metzker ML, Naylor SL, Kucherlapati RS, Nelson DL, Weinstock GM, Sakaki Y, Fujiyama A, Hattori M, Yada T, Toyoda A, Itoh T, Kawagoe C, Watanabe H, Totoki Y, Taylor T, Weissenbach J, Heilig R, Saurin W, Artiguenave F, Brottier P, Bruls T, Pelletier E, Robert C, Wincker P, Smith DR, Doucette-Stamm L, Rubenfield M, Weinstock K, Lee HM, Dubois J, Rosenthal A, Platzer M, Nyakatura G, Taudien S, Rump A, Yang H, Yu J, Wang J, Huang G, Gu J, Hood L, Rowen L, Madan A, Qin S, Davis RW, Federspiel NA, Abola AP, Proctor MJ, Myers RM, Schmutz J, Dickson M, Grimwood J, Cox DR, Olson MV, Kaul R, Shimizu N, Kawasaki K, Minoshima S, Evans GA, Athanasiou M, Schultz R, Roe BA, Chen F, Pan H, Ramser J, Lehrach H, Reinhardt R, McCombie WR, de la Bastide M, Dedhia N, Blocker H, Hornischer K, Nordsiek G, Agarwala R, Aravind L, Bailey JA, Bateman A, Batzoglou S, Birney E, Bork P, Brown DG, Burge CB, Cerutti L, Chen HC, Church D, Clamp M, Copley RR, Doerks T, Eddy SR, Eichler EE, Furey TS, Galagan J, Gilbert JG, Harmon C, Hayashizaki Y, Haussler D, Hermjakob H, Hokamp K, Jang W, Johnson LS, Jones TA, Kasif S, Kaspryzk A, Kennedy S, Kent WJ, Kitts P, Koonin EV, Korf I, Kulp D, Lancet D, Lowe TM, McLysaght A, Mikkelsen T, Moran JV, Mulder N, Pollara VJ, Ponting CP, Schuler G, Schultz J, Slater G, Smit AF, Stupka E, Szustakowski J, Thierry-Mieg D, Thierry-Mieg J, Wagner L, Wallis J, Wheeler R, Williams A, Wolf YI, Wolfe KH, Yang SP, Yeh RF, Collins F, Guyer MS, Peterson J, Felsenfeld A, Wetterstrand KA, Patrinos A, Morgan MJ, Szustakowki J, de Jong P, Catanese JJ, Osoegawa K, Shizuya H, Choi S, Chen YJ: Initial sequencing and analysis of the human genome. Nature. 2001, 409 (6822): 860-921. 10.1038/35057062.PubMedView ArticleGoogle Scholar
- Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, Gocayne JD, Amanatides P, Ballew RM, Huson DH, Wortman JR, Zhang Q, Kodira CD, Zheng XH, Chen L, Skupski M, Subramanian G, Thomas PD, Zhang J, Gabor Miklos GL, Nelson C, Broder S, Clark AG, Nadeau J, McKusick VA, Zinder N, Levine AJ, Roberts RJ, Simon M, Slayman C, Hunkapiller M, Bolanos R, Delcher A, Dew I, Fasulo D, Flanigan M, Florea L, Halpern A, Hannenhalli S, Kravitz S, Levy S, Mobarry C, Reinert K, Remington K, Abu-Threideh J, Beasley E, Biddick K, Bonazzi V, Brandon R, Cargill M, Chandramouliswaran I, Charlab R, Chaturvedi K, Deng Z, Di Francesco V, Dunn P, Eilbeck K, Evangelista C, Gabrielian AE, Gan W, Ge W, Gong F, Gu Z, Guan P, Heiman TJ, Higgins ME, Ji RR, Ke Z, Ketchum KA, Lai Z, Lei Y, Li Z, Li J, Liang Y, Lin X, Lu F, Merkulov GV, Milshina N, Moore HM, Naik AK, Narayan VA, Neelam B, Nusskern D, Rusch DB, Salzberg S, Shao W, Shue B, Sun J, Wang Z, Wang A, Wang X, Wang J, Wei M, Wides R, Xiao C, Yan C, Yao A, Ye J, Zhan M, Zhang W, Zhang H, Zhao Q, Zheng L, Zhong F, Zhong W, Zhu S, Zhao S, Gilbert D, Baumhueter S, Spier G, Carter C, Cravchik A, Woodage T, Ali F, An H, Awe A, Baldwin D, Baden H, Barnstead M, Barrow I, Beeson K, Busam D, Carver A, Center A, Cheng ML, Curry L, Danaher S, Davenport L, Desilets R, Dietz S, Dodson K, Doup L, Ferriera S, Garg N, Gluecksmann A, Hart B, Haynes J, Haynes C, Heiner C, Hladun S, Hostin D, Houck J, Howland T, Ibegwam C, Johnson J, Kalush F, Kline L, Koduru S, Love A, Mann F, May D, McCawley S, McIntosh T, McMullen I, Moy M, Moy L, Murphy B, Nelson K, Pfannkoch C, Pratts E, Puri V, Qureshi H, Reardon M, Rodriguez R, Rogers YH, Romblad D, Ruhfel B, Scott R, Sitter C, Smallwood M, Stewart E, Strong R, Suh E, Thomas R, Tint NN, Tse S, Vech C, Wang G, Wetter J, Williams S, Williams M, Windsor S, Winn-Deen E, Wolfe K, Zaveri J, Zaveri K, Abril JF, Guigo R, Campbell MJ, Sjolander KV, Karlak B, Kejariwal A, Mi H, Lazareva B, Hatton T, Narechania A, Diemer K, Muruganujan A, Guo N, Sato S, Bafna V, Istrail S, Lippert R, Schwartz R, Walenz B, Yooseph S, Allen D, Basu A, Baxendale J, Blick L, Caminha M, Carnes-Stine J, Caulk P, Chiang YH, Coyne M, Dahlke C, Mays A, Dombroski M, Donnelly M, Ely D, Esparham S, Fosler C, Gire H, Glanowski S, Glasser K, Glodek A, Gorokhov M, Graham K, Gropman B, Harris M, Heil J, Henderson S, Hoover J, Jennings D, Jordan C, Jordan J, Kasha J, Kagan L, Kraft C, Levitsky A, Lewis M, Liu X, Lopez J, Ma D, Majoros W, McDaniel J, Murphy S, Newman M, Nguyen T, Nguyen N, Nodell M, Pan S, Peck J, Peterson M, Rowe W, Sanders R, Scott J, Simpson M, Smith T, Sprague A, Stockwell T, Turner R, Venter E, Wang M, Wen M, Wu D, Wu M, Xia A, Zandieh A, Zhu X: The sequence of the human genome. Science. 2001, 291 (5507): 1304-1351. 10.1126/science.1058040.PubMedView ArticleGoogle Scholar
- Myers EW, Sutton GG, Delcher AL, Dew IM, Fasulo DP, Flanigan MJ, Kravitz SA, Mobarry CM, Reinert KH, Remington KA, Anson EL, Bolanos RA, Chou HH, Jordan CM, Halpern AL, Lonardi S, Beasley EM, Brandon RC, Chen L, Dunn PJ, Lai Z, Liang Y, Nusskern DR, Zhan M, Zhang Q, Zheng X, Rubin GM, Adams MD, Venter JC: A whole-genome assembly of Drosophila. Science. 2000, 287 (5461): 2196-2204. 10.1126/science.287.5461.2196.PubMedView ArticleGoogle Scholar
- Gibbs RA, Weinstock GM, Metzker ML, Muzny DM, Sodergren EJ, Scherer S, Scott G, Steffen D, Worley KC, Burch PE, Okwuonu G, Hines S, Lewis L, DeRamo C, Delgado O, Dugan-Rocha S, Miner G, Morgan M, Hawes A, Gill R, Celera, Holt RA, Adams MD, Amanatides PG, Baden-Tillson H, Barnstead M, Chin S, Evans CA, Ferriera S, Fosler C, Glodek A, Gu Z, Jennings D, Kraft CL, Nguyen T, Pfannkoch CM, Sitter C, Sutton GG, Venter JC, Woodage T, Smith D, Lee HM, Gustafson E, Cahill P, Kana A, Doucette-Stamm L, Weinstock K, Fechtel K, Weiss RB, Dunn DM, Green ED, Blakesley RW, Bouffard GG, De Jong PJ, Osoegawa K, Zhu B, Marra M, Schein J, Bosdet I, Fjell C, Jones S, Krzywinski M, Mathewson C, Siddiqui A, Wye N, McPherson J, Zhao S, Fraser CM, Shetty J, Shatsman S, Geer K, Chen Y, Abramzon S, Nierman WC, Havlak PH, Chen R, Durbin KJ, Egan A, Ren Y, Song XZ, Li B, Liu Y, Qin X, Cawley S, Worley KC, Cooney AJ, D'Souza LM, Martin K, Wu JQ, Gonzalez-Garay ML, Jackson AR, Kalafus KJ, McLeod MP, Milosavljevic A, Virk D, Volkov A, Wheeler DA, Zhang Z, Bailey JA, Eichler EE, Tuzun E, Birney E, Mongin E, Ureta-Vidal A, Woodwark C, Zdobnov E, Bork P, Suyama M, Torrents D, Alexandersson M, Trask BJ, Young JM, Huang H, Wang H, Xing H, Daniels S, Gietzen D, Schmidt J, Stevens K, Vitt U, Wingrove J, Camara F, Mar Alba M, Abril JF, Guigo R, Smit A, Dubchak I, Rubin EM, Couronne O, Poliakov A, Hubner N, Ganten D, Goesele C, Hummel O, Kreitler T, Lee YA, Monti J, Schulz H, Zimdahl H, Himmelbauer H, Lehrach H, Jacob HJ, Bromberg S, Gullings-Handley J, Jensen-Seaman MI, Kwitek AE, Lazar J, Pasko D, Tonellato PJ, Twigger S, Ponting CP, Duarte JM, Rice S, Goodstadt L, Beatson SA, Emes RD, Winter EE, Webber C, Brandt P, Nyakatura G, Adetobi M, Chiaromonte F, Elnitski L, Eswara P, Hardison RC, Hou M, Kolbe D, Makova K, Miller W, Nekrutenko A, Riemer C, Schwartz S, Taylor J, Yang S, Zhang Y, Lindpaintner K, Andrews TD, Caccamo M, Clamp M, Clarke L, Curwen V, Durbin R, Eyras E, Searle SM, Cooper GM, Batzoglou S, Brudno M, Sidow A, Stone EA, Venter JC, Payseur BA, Bourque G, Lopez-Otin C, Puente XS, Chakrabarti K, Chatterji S, Dewey C, Pachter L, Bray N, Yap VB, Caspi A, Tesler G, Pevzner PA, Haussler D, Roskin KM, Baertsch R, Clawson H, Furey TS, Hinrichs AS, Karolchik D, Kent WJ, Rosenbloom KR, Trumbower H, Weirauch M, Cooper DN, Stenson PD, Ma B, Brent M, Arumugam M, Shteynberg D, Copley RR, Taylor MS, Riethman H, Mudunuri U, Peterson J, Guyer M, Felsenfeld A, Old S, Mockrin S, Collins F: Genome sequence of the Brown Norway rat yields insights into mammalian evolution. Nature. 2004, 428 (6982): 493-521. 10.1038/nature02426.PubMedView ArticleGoogle Scholar
- Initial sequence of the chimpanzee genome and comparison with the human genome. Nature. 2005, 437 (7055): 69-87. 10.1038/nature04072.
- Ensembl: Ensembl. [http://www.ensembl.org]
- UCSCBrowser: UCSC Browser. [http://genome.ucsc.edu/]
- NCBI_Browser: NCBI Browser. [http://www.ncbi.nlm.nih.gov/mapview/]
- de Bruijn B, Martin J: Getting to the (c)ore of knowledge: mining biomedical literature. Int J Med Inform. 2002, 67 (1-3): 7-18. 10.1016/S1386-5056(02)00050-3.PubMedView ArticleGoogle Scholar
- Muller HM, Kenny EE, Sternberg PW: Textpresso: an ontology-based information retrieval and extraction system for biological literature. PLoS Biol. 2004, 2 (11): e309-10.1371/journal.pbio.0020309.PubMed CentralPubMedView ArticleGoogle Scholar
- Jenssen TK, Laegreid A, Komorowski J, Hovig E: A literature network of human genes for high-throughput analysis of gene expression. Nat Genet. 2001, 28 (1): 21-28. 10.1038/88213.PubMedGoogle Scholar
- Alako BT, Veldhoven A, van Baal S, Jelier R, Verhoeven S, Rullmann T, Polman J, Jenster G: CoPub Mapper: mining MEDLINE based on search term co-publication. BMC Bioinformatics. 2005, 6 (1): 51-10.1186/1471-2105-6-51.PubMed CentralPubMedView ArticleGoogle Scholar
- LaBaer J: Mining the literature and large datasets. Nat Biotechnol. 2003, 21 (9): 976-977. 10.1038/nbt0903-976b.PubMedView ArticleGoogle Scholar
- Rubinstein R, Simon I: MILANO--custom annotation of microarray results using automatic literature searches. BMC Bioinformatics. 2005, 6: 12-10.1186/1471-2105-6-12.PubMed CentralPubMedView ArticleGoogle Scholar
- Delong M, Yao G, Wang Q, Dobra A, Black EP, Chang JT, Bild A, West M, Nevins JR, Dressman H: DIG--a system for gene annotation and functional discovery. Bioinformatics. 2005, 21 (13): 2957-2959. 10.1093/bioinformatics/bti467.PubMedView ArticleGoogle Scholar
- Majumder PK, Yeh JJ, George DJ, Febbo PG, Kum J, Xue Q, Bikoff R, Ma H, Kantoff PW, Golub TR, Loda M, Sellers WR: Prostate intraepithelial neoplasia induced by prostate restricted Akt activation: the MPAKT model. Proc Natl Acad Sci U S A. 2003, 100 (13): 7841-7846. 10.1073/pnas.1232229100.PubMed CentralPubMedView ArticleGoogle Scholar
- Majumder PK, Febbo PG, Bikoff R, Berger R, Xue Q, McMahon LM, Manola J, Brugarolas J, McDonnell TJ, Golub TR, Loda M, Lane HA, Sellers WR: mTOR inhibition reverses Akt-dependent prostate intraepithelial neoplasia through regulation of apoptotic and HIF-1-dependent pathways. Nat Med. 2004, 10 (6): 594-601. 10.1038/nm1052.PubMedView ArticleGoogle Scholar
- Loda M, Capodieci P, Mishra R, Yao H, Corless C, Grigioni W, Wang Y, Magi-Galluzzi C, Stork PJ: Expression of mitogen-activated protein kinase phosphatase-1 in the early phases of human epithelial carcinogenesis. Am J Pathol. 1996, 149 (5): 1553-1564.PubMed CentralPubMedGoogle Scholar
- Rossi S, Graner E, Febbo P, Weinstein L, Bhattacharya N, Onody T, Bubley G, Balk S, Loda M: Fatty acid synthase expression defines distinct molecular signatures in prostate cancer. Mol Cancer Res. 2003, 1 (10): 707-715.PubMedGoogle Scholar
- MeSHer: MeSHer. [http://compbio.dfci.harvard.edu/mesher.html]
- GOMiner: GOMiner. [http://discover.nci.nih.gov/gominer/]
- Microsystems S: Java Software. [http://java.sun.com]
- Eclipse: Eclipse Software. [http://www.eclipse.org]
- Utilities NCBIP: NCBI Programming Utilities. [http://eutils.ncbi.nlm.nih.gov/entrez/query/static/eutils_help.html]
- Classes XMLB: XML Binding Classes. [http://java.sun.com/developer/technicalArticles/WebServices/jaxb]
- Classes APOI: Apachi POI Classes. [http://poi.apache.org]
- database SSOURCE: Stanford SOURCE. [http://genome-www5.stanford.edu/cgi-bin/source/sourceSearch]
- Schijvenaars BJ, Mons B, Weeber M, Schuemie MJ, van Mulligen EM, Wain HM, Kors JA: Thesaurus-based disambiguation of gene symbols. BMC Bioinformatics. 2005, 6: 149-10.1186/1471-2105-6-149.PubMed CentralPubMedView ArticleGoogle Scholar
- Biocarta: Biocarta. [http://www.biocarta.com/genes/allPathways.asp]
- Stegmaier K, Ross KN, Colavito SA, O'Malley S, Stockwell BR, Golub TR: Gene expression-based high-throughput screening(GE-HTS) and application to leukemia differentiation. Nat Genet. 2004, 36 (3): 257-263. 10.1038/ng1305.PubMedView ArticleGoogle Scholar
- van de Vijver MJ, He YD, van't Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, Parrish M, Atsma D, Witteveen A, Glas A, Delahaye L, van der Velde T, Bartelink H, Rodenhuis S, Rutgers ET, Friend SH, Bernards R: A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med. 2002, 347 (25): 1999-2009. 10.1056/NEJMoa021967.PubMedView ArticleGoogle Scholar
- van 't Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R, Friend SH: Gene expression profiling predicts clinical outcome of breast cancer. Nature. 2002, 415 (6871): 530-536. 10.1038/415530a.PubMedView ArticleGoogle Scholar
- Tallman MS, Andersen JW, Schiffer CA, Appelbaum FR, Feusner JH, Ogden A, Shepherd L, Willman C, Bloomfield CD, Rowe JM, Wiernik PH: All-trans-retinoic acid in acute promyelocytic leukemia. N Engl J Med. 1997, 337 (15): 1021-1028. 10.1056/NEJM199710093371501.PubMedView ArticleGoogle Scholar
- Schopf RE, Mattar J, Meyenburg W, Scheiner O, Hammann KP, Lemmel EM: Measurement of the respiratory burst in human monocytes and polymorphonuclear leukocytes by nitro blue tetrazolium reduction and chemiluminescence. J Immunol Methods. 1984, 67 (1): 109-117. 10.1016/0022-1759(84)90090-5.PubMedView ArticleGoogle Scholar
- Duffy MJ, Maguire TM, Hill A, McDermott E, O'Higgins N: Metalloproteinases: role in breast carcinogenesis, invasion and metastasis. Breast Cancer Res. 2000, 2 (4): 252-257. 10.1186/bcr65.PubMed CentralPubMedView ArticleGoogle Scholar
- Gupta GP, Nguyen DX, Chiang AC, Bos PD, Kim JY, Nadal C, Gomis RR, Manova-Todorova K, Massague J: Mediators of vascular remodelling co-opted for sequential steps in lung metastasis. Nature. 2007, 446 (7137): 765-770. 10.1038/nature05760.PubMedView ArticleGoogle Scholar
- Sledge GW: Vascular endothelial growth factor in breast cancer: biologic and therapeutic aspects. Semin Oncol. 2002, 29 (3 Suppl 11): 104-110. 10.1053/sonc.2002.34062.PubMedView ArticleGoogle Scholar
- GeneCite: GeneCite.
- Stuart JM, Segal E, Koller D, Kim SK: A gene-coexpression network for global discovery of conserved genetic modules. Science. 2003, 302 (5643): 249-255. 10.1126/science.1087447.PubMedView ArticleGoogle Scholar
- Bild A, Febbo PG: Application of a priori established gene sets to discover biologically important differential expression in microarray data. Proc Natl Acad Sci U S A. 2005, 102 (43): 15278-15279. 10.1073/pnas.0507477102.PubMed CentralPubMedView ArticleGoogle Scholar
- Bild AH, Yao G, Chang JT, Wang Q, Potti A, Chasse D, Joshi MB, Harpole D, Lancaster JM, Berchuck A, Olson JA, Marks JR, Dressman HK, West M, Nevins JR: Oncogenic pathway signatures in human cancers as a guide to targeted therapies. Nature. 2005Google Scholar
- Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP: Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005, 102 (43): 15545-15550. 10.1073/pnas.0506580102.PubMed CentralPubMedView ArticleGoogle Scholar
- Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, Puigserver P, Carlsson E, Ridderstrale M, Laurila E, Houstis N, Daly MJ, Patterson N, Mesirov JP, Golub TR, Tamayo P, Spiegelman B, Lander ES, Hirschhorn JN, Altshuler D, Groop LC: PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003, 34 (3): 267-273. 10.1038/ng1180.PubMedView ArticleGoogle Scholar
- Zeeberg BR, Feng W, Wang G, Wang MD, Fojo AT, Sunshine M, Narasimhan S, Kane DW, Reinhold WC, Lababidi S, Bussey KJ, Riss J, Barrett JC, Weinstein JN: GoMiner: a resource for biological interpretation of genomic and proteomic data. Genome Biol. 2003, 4 (4): R28-10.1186/gb-2003-4-4-r28.PubMed CentralPubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.